This hits #RHEL 9, #AlmaLinux 9, #Rocky 9, #CentOS Stream 9: all the same under the hood, all the same headache.
Check your limits.conf BEFORE your next upgrade. You'll thank yourself later.
#NMR #Bruker #TopSpin #ScientificComputing #LabIT

The fix is annoyingly simple once you know:

Add to /etc/security/limits.conf:

* soft memlock unlimited
* hard memlock unlimited

Reboot. That's it. You're done.

What happened: #RHEL9 slashed the memory locking limit to a measly 8 MB in the name of "security hardening." Real-time scientific software needs way more than that to function. Shimming your magnet? Not anymore.

Anyone looking into it, here’s a useful resource: wiki.serversatho.me/nextcloud

Now on the ESP-IDF Component Registry:

idf.py add-dependency "debsahu/espidf-nats^1.0.2"

Header-only, works with ESP-IDF 4.4+, supports ESP32/S2/S3/C3.

PRs welcome!

github.com/debsahu/espidf-nats

The Claude workflow was wild - I'd describe a NATS feature, Claude would implement it, we'd iterate on edge cases.

Went from ~200 lines to 2000+ lines of header-only C++. All the JetStream API complexity handled in clean callbacks.

Perfect for IoT:

• KV Store → device config, sensor calibration, device twins
• Object Store → OTA firmware updates, log uploads
• JetStream → guaranteed delivery for critical sensor data
• Offline buffering → handles spotty WiFi gracefully

More reliability features:

• NATS 2.0 Headers (HPUB/HMSG)
• Request timeouts with callbacks
• Async non-blocking connect
• Message buffering when offline
• Exponential backoff reconnect
• Connection metrics & error codes

What Claude helped me build:

• TLS/SSL with mTLS support
• Full JetStream - streams, consumers, pull-based delivery, ACKs
• Key-Value Store for device config
• Object Store for firmware/large files
• Multi-server failover

Sure, here’s what it ended up as. TrueNAS scale offline backup of my home TrueNAS scale.

People are upset about Opus weekly limits, and tone deaf Anthropic is saying this:

Yes that’s was my second choice!